In testing environments, tests can often involve multiple pieces of hardware running for different periods, all of which consume variable amounts of power and output heat. While it is possible to use the manufacturer's specifications to model heat and power usage, this does not reflect accurately the behavior over a specific workload.
If the test environment is automated, tests may be scheduled which do not produce an ideal mix of temperature output and power usage, which may lead to spikes in both temperature and power usage. This may conflict with the aims of data centre management, where it is preferable to smooth both power usage and temperature output to allow for most efficient management of environmental factors and cost.
In increasingly resource constrained environments, organisations may need to make decisions about the effectiveness of tests by comparing their environmental impact versus the benefit in exposing defects.
Therefore, there is a need in the art to address the aforementioned problems.